AMD CEO Lisa Su visited Samsung on Oct 7 to discuss HBM supply and foundry orders. Samsung seeks foundry business in exchange, but AMD officials cited high risk in outsourcing to Samsung. Neither party has confirmed the deal.
Samsung is reportedly planning to exit the automotive application processor (AP) business it has pursued since 2018. According to the report citing industry sources, once Samsung completes production under existing supply commitments to Hyundai Motor (agreement secured in 2023) and Germany's BMW (secured last year), it plans to discontinue the development and mass production of successor Exynos Auto products; this is a reported plan, not a confirmed halt. Samsung will focus its System LSI operations on image sensors and next-generation mobile APs as it seeks to return the division to full-year profit next year. The stated causes are high development costs from stringent automotive safety and durability requirements and Qualcomm's effective dominance of the automotive AP market, which makes it difficult for Samsung to gain share. The System LSI and foundry businesses have together posted annual losses totaling trillions of won. In its latest detailed update released with second-quarter results in July, Samsung said expanded sales of mobile SoCs and image sensors helped System LSI post record first-half revenue in 2026. Samsung is developing the Exynos 2700 mobile AP, scheduled to enter mass production in the fourth quarter on Samsung Foundry's second-generation 2nm process, and seeks to expand its adoption in the Galaxy S27 series due in 2027; Samsung spent KRW 7.44 trillion on mobile APs from external suppliers in the first half of 2026. Samsung also plans to expand businesses including display driver ICs (DDIs), power management ICs (PMICs), and chip design for data centers.
Micron and Samsung are shifting HBM base die production to TSMC, with Micron projecting higher ASP and margins for NVHBM. This follows NVIDIA's NVHBM architecture which increases base die value. Bullish: Outsourcing to TSMC and higher ASP expectations for custom HBM.
Infineon Technologies AG (FSE: IFX / OTCQX: IFNNY) and ZuriQ AG, a Swiss quantum computing company spun out of ETH Zurich, today announced the expansion of their collaboration to advance scalable quantum computing hardware. The partnership combines ZuriQ's trapped-ion quantum computing architecture with Infineon's expertise in semiconductor manufacturing, advanced packaging, and integrated photonics. Previous development work demonstrated feasibility of the architecture by trapping a 3×3 array of nine individually controlled ions, achieving the largest two-dimensional array of its kind to date; the next step is to scale the technology to significantly larger qubit counts. ZuriQ has developed a natively two-dimensional Penning micro-trap architecture that uses electric and magnetic fields to move ions directly across the chip, avoiding complex junction structures and making it easier to accommodate larger qubit arrays. This is a planned expansion of collaboration and has not been shown to involve any price movement or capacity figures.
This is a technical commentary article by Bhanu Pandey, a product engineer for Calibre Design Solutions at Siemens Digital Industries Software, addressing the 'floating net' structural defect in analog and mixed-signal (AMS) verification. Citing the most recent Wilson Research Group / Siemens EDA Functional Verification Study, the article states that first-silicon success rates have hit a historic low, with nearly 68% of designs requiring two or more spins, and that AMS verification remains a primary source of these re-spins. It explains how floating nets arise (a node disconnected from all low-impedance paths in power-down modes, its voltage determined by trapped charge on parasitic capacitances, drifting over time due to leakage currents, potentially causing latch-up or permanent physical damage) and lists three reasons SPICE simulation misses the defect: convergence artifacts (gmin stepping) that mask it, the time-scale problem, and testbench coverage limits. A case study of a design with an internally labeled 'VICTIM' net shows the defect was masked in simulation but in silicon caused a massive current spike that damaged the output stage, leading to a costly field failure and a subsequent re-spin. The article advocates automated structural analysis (e.g., the Calibre Insight Analyzer) to detect floating nets before fabrication and shift reliability verification left. It contains no pricing, supply-demand, capacity or procurement information and is background technical content.
According to a report by Chosun.biz (published 8 October 2026), AMD CEO Lisa Su is visiting South Korea to secure additional supply of HBM memory for AI accelerators, saying 'We are going to talk about our broader partnership' with Samsung executives; Samsung, meanwhile, reportedly 'hopes to turn its memory supply relationship with AMD into foundry' orders for logic chips. The visit comes a little less than seven months after the two companies signed a non-binding memorandum of understanding (MOU) this March, effective for three years and covering four areas: supply of next-generation memory products including HBM4 and advanced DDR5; co-development and optimization of memory solutions for AI and high-performance computing systems; exploration of foundry manufacturing opportunities for future AMD platforms; and co-optimization of advanced packaging and related semiconductor technologies. Samsung has become a principal supplier of HBM4 for AMD's Instinct MI400-series accelerators; expanding collaboration beyond memory supply is considerably more complicated, as it involves considerable efforts to design chips for Samsung Foundry's process technologies and risks associated with a product ramp at a new contract chipmaker — AMD's most recent Annual Report states that transitioning production of some products to new manufacturers may cause delayed product introductions, lower yields or poorer performance. AMD has already started shipments of its Instinct MI455 accelerators with 432GB of HBM4 onboard as well as rack-scale Helios AI systems to interested parties; Su said 'Samsung Electronics and SK hynix are key supply-chain partners for AMD,' that 'AI demand is extremely high' so AMD is planning its long-term supply chain with partners, and that 'It is important to plan three to five years ahead with our partners.' The article concludes that AMD is negotiating for more memory supply and whether it will take the risk of adopting Samsung Foundry beyond niche programmable chips 'only time will tell.'
The industry is shifting to mmWave frequencies above 30GHz and adopting waveguide technologies to address signal loss. Traditional interconnects are becoming obsolete, prompting a transition to GaN as a viable alternative to GaAs in high-frequency applications.
Rohm Semiconductor is expanding its back-end semiconductor operations outsourcing and growing its R&D footprint in India, while retaining wafer fabrication in-house. The source does not state the specific scale, partners, or cause of the outsourcing expansion.
Infineon and ZuriQ plan to expand their collaboration to scale qubits; they have demonstrated the feasibility of the architecture by trapping a 3×3 ion array, the largest two-dimensional array of its kind to date; the next step is to scale the technology to significantly larger qubit counts.
Hanmi Semiconductor secured a KRW24.48 billion (approximately US$18.3 million) equipment order from Samsung Electro-Mechanics, a rare deal with a Samsung Group affiliate after years of limited business following a patent dispute. The contract was signed a week after Samsung Electro-Mechanics announced a KRW6.78 trillion expansion of its AI server substrate business, as demand for advanced chip packaging continues to grow.
India's outsourced semiconductor assembly and test (OSAT) sector has moved from inaugurations to its first export shipments, and the challenge now is scale. According to the report, the new plants are small and still rely on foreign engineers and partners, and the government's second subsidy round offers less support for conventional packaging than the first. Japanese chipmakers are among the earliest customers, but they are spreading work across several Indian partners rather than building their own plants.
Microsoft released its first Surface notebook powered by Nvidia's RTX Spark SoC (the N1X chip), the Surface Laptop Ultra, during an event in San Francisco on Wednesday. It is one of several new Copilot+ PCs aimed at 'builders,' available for preorder starting today and expected to begin shipping next week. It starts at $2,599 for the gray variant with the lower-end 18-core N1X, 24 GB of unified memory, and a 512 GB user-serviceable SSD; stepping up to the 20-core chip, 128 GB of memory, and 1 TB of storage costs $5,899. Microsoft also announced a mini PC style development kit based on the N1X priced at $5,999. The machine ships with either an 18- or 20-core CPU designed by MediaTek and Arm and an Nvidia Blackwell-based RTX GPU with between 5,120 and 6,144 cores, with up to 128 GB of memory. The source does not state any impact on supply or pricing.
Samsung Electronics reported a record 783% YoY operating profit surge to 107.4 trillion KRW in Q3 2026, driven by HBM4 demand and memory price gains. HBM bit shipments jumped nearly 50% QoQ, while DRAM prices rose amid tight supply. Bullish: Supply constraints and price uptrends are confirmed.
Rambus Inc. issued a press announcement ahead of the recent AI Infra Summit and was present at the summit to spotlight its new production-ready CryptoManager Root of Trust supporting the Caliptra specification. The solution introduces hardware-level security orchestration built directly around the unmodified open-source Caliptra core; by integrating a secure RISC-V processor, protected memory, and dedicated cryptographic accelerators alongside specialized drivers, it expands security visibility across the entire System-on-Chip (SoC). The stated cause is the gap between an open-source specification and production-ready, commercially deployable silicon: an open core typically lacks platform-wide security orchestration, and chip vendors otherwise shoulder all integration and schedule risks alone. The offering covers physical attack defense (side-channel analysis and fault injection countermeasures), cryptographic agility blending classical algorithms like AES, SHA, and RSA with emerging Post-Quantum Cryptography standards such as ML-KEM and ML-DSA, and readiness for certifications including FIPS 140-3, PSA, or SESIP.
According to STAR Market Daily, semiconductor testing demand is surging as increasingly complex AI chips require more intensive processes, tightening available capacity; OSAT provider Sigurd Microelectronics has posted monthly revenue above NT$2 billion for three consecutive months, and customers are increasingly signing long-term agreements to secure testing capacity — most have signed capacity agreements lasting more than a year, others have reserved capacity through dedicated equipment, and some memory customers are even willing to pay premiums to lock in capacity. On pricing, Sigurd said it could not speak for the industry but confirmed that it has already begun raising its own prices. Per TechNews, AI, CPUs, GPUs, ASICs, and silicon photonics generated about NT$3.7 billion in revenue for Sigurd in the first nine months of 2026, up from roughly NT$2.5 billion a year earlier, with their share of total revenue rising from 20% to 24%, and the company expects the share to exceed 30% in 2027. Silicon photonics is also a key growth driver: Sigurd has expanded its optical communications testing capabilities to cover EICs, PICs, and optical engines, spanning testing technology, R&D engineering, and volume production. Strong demand is accelerating capacity expansion: Sigurd's second Hukou plant began production in July, with about 40% of capacity online as of October and already locked in by customers; Zhongxing Plant 3 is scheduled for completion in the fourth quarter of 2026 and volume production in the first quarter of 2027. Sigurd has raised its 2026 capex to NT$8.8 billion, 48% above its original plan, focused on advanced testing equipment, and is in discussions with several key customers and could raise capex again in the coming months. As background: per Seoul Economic Daily citing Nomura Securities, NVIDIA Blackwell's final test time is estimated at four times that of Hopper, while Rubin could require seven times as long, with testing's share of GPU manufacturing costs projected to rise from 1.9% for Hopper to 2.5% for Blackwell and 3.3% for Rubin; per ETNews, Digital Frontier and UniTest have secured a combined KRW 283.94 billion in HBM4 wafer tester orders from SK hynix this year, and as HBM complexity increases, more test items and longer testing times could drive testing equipment demand to grow faster than HBM output itself.
Micron's Taoyuan union in Taiwan secured the legal right to strike via a vote held between October 1 and 6, in which 1,994 of 2,258 registered members voted in favor, representing 88.3% of total membership and exceeding the required majority; Taoyuan is the first of Micron's separate Taoyuan and Taichung enterprise unions to secure the right to strike. The Taichung union is scheduled for a second labor-management mediation session on October 22 and could proceed to a strike vote if no agreement is reached. Union chairman Lin Che-jui called on Micron to present concrete proposals and negotiate as the company's board meets on October 8 and 9; without meaningful progress the union will discuss further action, and members will be notified by text message once the timing of any surprise strike is confirmed. The union is also planning a demonstration on Ketagalan Boulevard on October 19, encouraging employees to take leave to participate. Its demands include a profit-sharing system tied to 15% of Micron's operating profit: an additional one-time FY2026 bonus based on 15% of global operating profit, estimated at about 83 months of pay per Taiwan employee; an overhaul of the IPP system from FY2027 to distribute 15% of operating profit quarterly; plus 'no free-rider' and 'non-retaliation' clauses. The backdrop cited is compensation disparity: in 2025, Micron's global employee IPP performance factor was 93.9%, translating into about 2.3 months of performance bonuses for Taiwan employees before being adjusted to 2.6 months, while the CEO received annual cash performance compensation equivalent to 32.4 months of pay; only about 40% of Micron Taiwan employees had previously received stock awards. TrendForce notes that Micron's Taoyuan fab primarily produces DRAM; if a strike takes place and disrupts equipment maintenance and process monitoring by engineers, production yields and stability could decline, potentially affecting output, and the impact on supply could gradually increase if the strike is prolonged; however, the actual impact will depend on the scale and duration of the strike as well as Micron's contingency measures, and the outcome of labor-management negotiations still requires close attention.
Ambarella's N1-655 chip can run LLMs and VLMs on the edge, supporting low-power edge AI systems for autonomous robots. The source only describes the product's stated capability and provides no pricing, production, shipment or customer information, and the source does not state the cause.
imec expands its IC-Link collaboration with TSMC to provide managed cloud services for EDA flows and advanced packaging support via the 3DFabric Alliance.
Siemens has again sponsored a Wilson-led survey on the state of (functional) verification in the industry, and Harry Foster (Chief Scientist, Verification at Siemens) has condensed his major takeaways from the report. The survey shows an apparent revolutionary change in first-silicon trends: where an average of ~30% of designs claimed first silicon success over at least a decade, this metric has been dropping, with the latest survey now showing 5%. The analysis notes that silicon respins are expensive and risk missing market windows, and Foster breaks the trend down into 5 primary factors, including growing accelerator-based volumes (coherency issues in AI/GPU and CPU cluster designs), the increasingly intimate link between software and hardware, a changing verification language mix (System Verilog now neck and neck with C/C++ and Python, with PSS growing fast), and significant survey-reported growth in respins around firmware, safety and security. Foster does not see the language mix as a contributor to the low success rate, viewing it instead as a signal of how verification teams are responding to complexity.
SK hynix faces HBM hybrid bonding challenges, delaying sample production behind Samsung. This threatens its market leadership and risks losing high-capacity customers. Bearish for SK hynix supply chain competitiveness.
Whose chips are really inside the hottest new products? We open them up and show you, one part at a time.
| Analog | $0.68 | −5.9% |
|---|---|---|
| Connector | $0.75 | +0.2% |
| CPU | $386.74 | −21.3% |
| DDR4 | $88.40 | −0.2% |
| DDR52026-W13 | $14.89 | −11.2% |
| GPU | $478.20 | +8.2% |
| MCU | $0.68 | −61.7%Unusual |
| MLCC | $0.01 | +54.0%Unusual |
| MOSFET | $0.12 | −82.8%Unusual |
| PMIC2026-W13 | $0.44 | −12.5% |
| SSD/HDD | $288.43 | n/a |